SecureGS: Boosting the Security and Fidelity of 3D Gaussian Splatting Steganography

2025-03-08

AI Summary

3D Gaussian Splatting (3DGS)-based steganography suffers from weak geometric structural security, degraded rendering fidelity, and high computational overhead in privacy-preserving applications. Method: We propose a Hybrid Decoupled Gaussian Encryption framework. It introduces a novel density-aware dynamic anchor growth-and-pruning strategy to ensure geometric-level security and adaptive localization of hidden regions. Integrating Scaffold-GS anchor design with neural decoding, the framework supports joint embedding of offset, scale, rotation, and RGB attributes. Additionally, we incorporate a privacy-preserving neural network and adaptive anchor optimization. Results: Experiments demonstrate that our method consistently outperforms state-of-the-art approaches in PSNR/SSIM fidelity, inference speed, and robustness against statistical analysis—reducing empirical attack success rates to below 2%.

Technology Category

• Machine Learning: Privacy
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Security and Privacy: Applications of cryptography
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

3D Gaussian Splatting (3DGS) has emerged as a premier method for 3D representation due to its real-time rendering and high-quality outputs, underscoring the critical need to protect the privacy of 3D assets. Traditional NeRF steganography methods fail to address the explicit nature of 3DGS since its point cloud files are publicly accessible. Existing GS steganography solutions mitigate some issues but still struggle with reduced rendering fidelity, increased computational demands, and security flaws, especially in the security of the geometric structure of the visualized point cloud. To address these demands, we propose a SecureGS, a secure and efficient 3DGS steganography framework inspired by Scaffold-GS's anchor point design and neural decoding. SecureGS uses a hybrid decoupled Gaussian encryption mechanism to embed offsets, scales, rotations, and RGB attributes of the hidden 3D Gaussian points in anchor point features, retrievable only by authorized users through privacy-preserving neural networks. To further enhance security, we propose a density region-aware anchor growing and pruning strategy that adaptively locates optimal hiding regions without exposing hidden information. Extensive experiments show that SecureGS significantly surpasses existing GS steganography methods in rendering fidelity, speed, and security.

Problem

• Enhance security and fidelity in 3D Gaussian Splatting steganography.
• Address vulnerabilities in existing 3DGS steganography methods.
• Protect privacy of 3D assets with efficient encryption mechanisms.

Innovation

• Hybrid decoupled Gaussian encryption mechanism
• Density region-aware anchor growing strategy
• Privacy-preserving neural networks for retrieval